Cory Doctorow on the Big AI Lie: 'bad at sandboxes' vs 'accidentally built God'
AlexTensor · x · 2026-09-24
Cory Doctorow calls out a contradictory pair of claims from AI companies: admitting "you suck at security sandboxes" while simultaneously suggesting they "accidentally built God" that might turn us into paper clips.
His point: if models are truly capable enough to threaten civilization, incompetent safeguards are catastrophic; if the failures are mere engineering glitches, the doomsday narrative is inflated marketing.
Related event: Cory Doctorow Debunks the 'Big AI Lie'(4 posts)→
More from AGI Musings
- New paper: 'Embodied Hijack' explains why users keep anthropomorphizing disembodied LLMs — MacrinePhD · 2026-09-24
- Code is no longer scarce, and product orgs haven't caught up — aakashgupta · 2026-09-24
- Survey of 1,534 knowledge workers: technical skill ranks 2nd for AI co-workers, 20th for humans — rseroter · 2026-09-24
- Sales rep predicts CROs will shrink sales teams within 24 months as AI boosts top performers — thedealdirector · 2026-09-24
- TheZvi Weekly: Opus 5.5 Lands to Strong Reviews, OpenAI Ships Cheaper Sol and Luna, Anti-ASI Bill Introduced — TheZvi · 2026-09-24
- Doomers vs Accelerationists: Tom Friedman's "Cry Wolf Over AI" Column Sparks Backlash — basedjensen · 2026-09-24